Opportunistic Screening for Atrial Fibrillation on Routine Chest Computed Tomography.

PURPOSE: Quantitative biomarkers from chest computed tomography (CT) can facilitate the incidental detection of important diseases. Atrial fibrillation (AFib) substantially increases the risk for comorbid conditions including stroke. This study investigated the relationship between AFib status and left atrial enlargement (LAE) on CT. MATERIALS AND METHODS: A total of 500 consecutive patients who had undergone nongated chest CTs were included, and left atrium maximal axial cross-sectional area (LA-MACSA), left atrium anterior-posterior dimension (LA-AP), and vertebral body cross-sectional area (VB-Area) were measured. Height, weight, age, sex, and diagnosis of AFib were obtained from the medical record. Parametric statistical analyses and receiver operating characteristic curves were performed. Machine learning classifiers were run with clinical risk factors and LA measurements to predict patients with AFib. RESULTS: Eighty-five patients with a diagnosis of AFib were identified. Mean LA-MACSA and LA-AP were significantly larger in patients with AFib than in patients without AFib (28.63 vs. 20.53 cm 2 , P <0.000001; 4.34 vs. 3.5 cm, P <0.000001, respectively), both with area under the curves (AUCs) of 0.73. Multivariable logistic regression analysis including age, sex, and VB-Area with LA-MACSA improved the AUC for predicting AFib (AUC=0.77). An LA-MACSA threshold of 30 cm 2 demonstrated high specificity for AFib diagnosis at 92% and sensitivity of 48%, and LA-AP threshold at 4.5 cm demonstrated 90% specificity and 42% sensitivity. A Bayesian machine learning model using age, sex, height, body surface area, and LA-MACSA predicted AFib with an AUC of 0.743. CONCLUSIONS: LA-MACSA or LA-AP can be rapidly measured from routine chest CT, and when >30 cm 2 and >4.5 cm, respectively, are specific indicators to predict patients at increased risk for AFib.

Whole Exome Sequencing Identifies PHF14 Mutations in Neurocytoma and Predicts Responsivity to the PDGFR Inhibitor Sunitinib.

Neurocytomas are rare low-grade brain tumors predominantly affecting young adults, but their cellular origin and molecular pathogenesis is largely unknown. We previously reported a sellar neurocytoma that secreted excess arginine vasopressin causing syndrome of inappropriate anti-diuretic hormone (SIADH). Whole exome sequencing in 21 neurocytoma tumor tissues identified somatic mutations in the plant homeodomain finger protein 14 (PHF14) in 3/21 (14%) tumors. Of these mutations, two were missense mutations and 4 caused splicing site losses, resulting in PHF14 dysfunction. Employing shRNA-mediated knockdown and CRISPR/Cas9-based knockout approaches, we demonstrated that loss of PHF14 increased proliferation and colony formation in five different human, mouse and rat mesenchymal and differentiated cell lines. Additionally, we demonstrated that PHF14 depletion resulted in upregulation of platelet derived growth factor receptor-alpha (PDGFRα) mRNA and protein in neuroblastoma SHSY-5Y cells and led to increased sensitivity to treatment with the PDGFR inhibitor Sunitinib. Furthermore, in a neurocytoma primary culture harboring splicing loss PHF14 mutations, overexpression of wild-type PHF14 and sunitinib treatment inhibited cell proliferation. Nude mice, inoculated with PHF14 knockout SHSY-5Y cells developed earlier and larger tumors than control cell-inoculated mice and Sunitinib administration caused greater tumor suppression in mice harboring PHF-14 knockout than control SHSY-5Y cells. Altogether our studies identified mutations of PHF14 in 14% of neurocytomas, demonstrate it can serve as an alternative pathway for certain cancerous behavior, and suggest a potential role for Sunitinib treatment in some patients with residual/recurrent neurocytoma.

Artificial Intelligence (AI) for Lung Nodules, From the AJR Special Series on AI Applications.

Interest in artificial intelligence (AI) applications for lung nodules continues to grow among radiologists, particularly with the expanding eligibility criteria and clinical utilization of lung cancer screening CT. AI has been heavily investigated for detecting and characterizing lung nodules and for guiding prognostic assessment. AI tools have also been used for image postprocessing (e.g., rib suppression on radiography or vessel suppression on CT) and for noninterpretive aspects of reporting and workflow, including management of nodule follow-up. Despite growing interest in and rapid development of AI tools and FDA approval of AI tools for pulmonary nodule evaluation, integration into clinical practice has been limited. Challenges to clinical adoption have included concerns about generalizability, regulatory issues, technical hurdles in implementation, and human skepticism. Further validation of AI tools for clinical use and demonstration of benefit in terms of patient-oriented outcomes also are needed. This article provides an overview of potential applications of AI tools in the imaging evaluation of lung nodules and discusses the challenges faced by practices interested in clinical implementation of such tools.

Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning.

Radiological findings on chest X-ray (CXR) have shown to be essential for the proper management of COVID-19 patients as the maximum severity over the course of the disease is closely linked to the outcome. As such, evaluation of future severity from current CXR would be highly desirable. We trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from an open-source dataset (COVID-19 image data collection) and from a multi-institutional local ICU dataset. The data was grouped into pairs of sequential CXRs and were categorized into three categories: 'Worse', 'Stable', or 'Improved' on the basis of radiological evolution ascertained from images and reports. Classical machine-learning algorithms were trained on the deep learning extracted features to perform immediate severity evaluation and prediction of future radiological trajectory. Receiver operating characteristic analyses and Mann-Whitney tests were performed. Deep learning predictions between "Worse" and "Improved" outcome categories and for severity stratification were significantly different for three radiological signs and one diagnostic ('Consolidation', 'Lung Lesion', 'Pleural effusion' and 'Pneumonia'; all P < 0.05). Features from the first CXR of each pair could correctly predict the outcome category between 'Worse' and 'Improved' cases with a 0.81 (0.74-0.83 95% CI) AUC in the open-access dataset and with a 0.66 (0.67-0.64 95% CI) AUC in the ICU dataset. Features extracted from the CXR could predict disease severity with a 52.3% accuracy in a 4-way classification. Severity evaluation trained on the COVID-19 image data collection had good out-of-distribution generalization when testing on the local dataset, with 81.6% of intubated ICU patients being classified as critically ill, and the predicted severity was correlated with the clinical outcome with a 0.639 AUC. CXR deep learning features show promise for classifying disease severity and trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.

Deep learning of chest X-rays can predict mechanical ventilation outcome in ICU-admitted COVID-19 patients.

The COVID-19 pandemic repeatedly overwhelms healthcare systems capacity and forced the development and implementation of triage guidelines in ICU for scarce resources (e.g. mechanical ventilation). These guidelines were often based on known risk factors for COVID-19. It is proposed that image data, specifically bedside computed X-ray (CXR), provide additional predictive information on mortality following mechanical ventilation that can be incorporated in the guidelines. Deep transfer learning was used to extract convolutional features from a systematically collected, multi-institutional dataset of COVID-19 ICU patients. A model predicting outcome of mechanical ventilation (remission or mortality) was trained on the extracted features and compared to a model based on known, aggregated risk factors. The model reached a 0.702 area under the curve (95% CI 0.707-0.694) at predicting mechanical ventilation outcome from pre-intubation CXRs, higher than the risk factor model. Combining imaging data and risk factors increased model performance to 0.743 AUC (95% CI 0.746-0.732). Additionally, a post-hoc analysis showed an increase performance on high-quality than low-quality CXRs, suggesting that using only high-quality images would result in an even stronger model.

Inversion group (IG) fitting: A new T1 mapping method for modified look-locker inversion recovery (MOLLI) that allows arbitrary inversion groupings and rest periods (including no rest period).

PURPOSE: The Modified Look-Locker Inversion Recovery (MOLLI) technique is used for T1 mapping in the heart. However, a drawback of this technique is that it requires lengthy rest periods in between inversion groupings to allow for complete magnetization recovery. In this work, a new MOLLI fitting algorithm (inversion group [IG] fitting) is presented that allows for arbitrary combinations of inversion groupings and rest periods (including no rest period). THEORY AND METHODS: Conventional MOLLI algorithms use a three parameter fitting model. In IG fitting, the number of parameters is two plus the number of inversion groupings. This increased number of parameters permits any inversion grouping/rest period combination. Validation was performed through simulation, phantom, and in vivo experiments. RESULTS: IG fitting provided T1 values with less than 1% discrepancy across a range of inversion grouping/rest period combinations. By comparison, conventional three parameter fits exhibited up to 30% discrepancy for some combinations. The one drawback with IG fitting was a loss of precision-approximately 30% worse than the three parameter fits. CONCLUSION: IG fitting permits arbitrary inversion grouping/rest period combinations (including no rest period). The cost of the algorithm is a loss of precision relative to conventional three parameter fits. Magn Reson Med 75:2332-2340, 2016. © 2015 Wiley Periodicals, Inc.

Quantification of Myocardial Extracellular Volume Fraction with Cardiac MR Imaging in Thalassemia Major.

Purpose To quantify myocardial extracellular volume (ECV) by using cardiac magnetic resonance (MR) imaging in thalassemia major and to investigate the relationship between ECV and myocardial iron overload. Materials and Methods With institutional review board approval and informed consent, 30 patients with thalassemia major (mean age ± standard deviation, 34.6 years ± 9.5) and 10 healthy control subjects (mean age, 31.5 years ± 4.4) were prospectively recruited (clinicaltrials.gov identification number NCT02090699). Nineteen patients (63.3%) had prior myocardial iron overload (defined as midseptal T2* < 20 msec on any prior cardiac MR images). Cardiac MR imaging at 1.5 T included cine steady-state free precession for ventricular function, T2* for myocardial iron quantification, and unenhanced and contrast material-enhanced T1 mapping. ECV was calculated with input of the patient's hematocrit level. Peak systolic global longitudinal strain by means of speckle tracking was assessed with same-day transthoracic echocardiography. Statistical analysis included use of the two-sample t test, Fisher exact test, and Spearman correlation. Results Unenhanced T1 values were significantly lower in patients with prior myocardial iron overload than in control subjects (850.3 ± 115.1 vs 1006.3 ± 35.4, P < .001) and correlated strongly with T2* values (r = 0.874, P < .001). Patients with prior myocardial iron overload had higher ECV than did patients without iron overload (31.3% ± 2.8 vs 28.2% ± 3.4, P = .030) and healthy control subjects (27.0% ± 3.1, P = .003). There was no difference in ECV between patients without iron overload and control subjects (P = .647). ECV correlated with lowest historical T2* (r = -0.469, P = .010) but did not correlate significantly with left ventricular ejection fraction (r = -0.216, P = .252) or global longitudinal strain (r = -0.164, P = .423). Conclusion ECV is significantly increased in thalassemia major and is associated with myocardial iron overload. These abnormalities may potentially reflect diffuse interstitial myocardial fibrosis. (©) RSNA, 2015 Online supplemental material is available for this article.

Fat fraction bias correction using T1 estimates and flip angle mapping.

PURPOSE: To develop a new method of reducing T1 bias in proton density fat fraction (PDFF) measured with iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL). MATERIALS AND METHODS: PDFF maps reconstructed from high flip angle IDEAL measurements were simulated and acquired from phantoms and volunteer L4 vertebrae. T1 bias was corrected using a priori T1 values for water and fat, both with and without flip angle correction. Signal-to-noise ratio (SNR) maps were used to measure precision of the reconstructed PDFF maps. PDFF measurements acquired using small flip angles were then compared to both sets of corrected large flip angle measurements for accuracy and precision. RESULTS: Simulations show similar results in PDFF error between small flip angle measurements and corrected large flip angle measurements as long as T1 estimates were within one standard deviation from the true value. Compared to low flip angle measurements, phantom and in vivo measurements demonstrate better precision and accuracy in PDFF measurements if images were acquired at a high flip angle, with T1 bias corrected using T1 estimates and flip angle mapping. CONCLUSION: T1 bias correction of large flip angle acquisitions using estimated T1 values with flip angle mapping yields fat fraction measurements of similar accuracy and superior precision compared to low flip angle acquisitions.